On 28/01/14 00:20, Darren Tucker wrote:
On Tue, Jan 28, 2014 at 6:38 AM, Tom G. Christensen <tgc@xxxxxxxxxxxxxxx> wrote:
[...]
Could there be some timing/performance issue causing those failures?
It's possible. We try to not make the tests timing-sensitive where
possible but some of them are inherently racy.
The slowest machine I've run the tests on is either a 166MHz powerpc
or a 170 MHz sparc. On the sparcs at least, building openssl with gcc
-mv8 or better makes a huge difference because it enables the hardware
multiply, and DH does a lot of multiplies.
Good call.
I rebuilt openssl with sparcv8 on the Solaris 7 host, and now it can
complete the rekey test.
I guess there is a timeout somewhere that causes the test to fail if it
takes too long to do the math.
I'm not sure I can get much more performance out of the IRIX 5.3 host,
so if there is a way to avoid or extend this timeout that would be nice.
-tgc
_______________________________________________
openssh-unix-dev mailing list
openssh-unix-dev@xxxxxxxxxxx
https://lists.mindrot.org/mailman/listinfo/openssh-unix-dev
